Novel Human Herpesvirus 8 Subtype D Strains in Vanuatu, Melanesia

We show human herpesvirus 8 with diverse molecular subtype D variants to be highly endemic among the Ni-Vanuatu population. Most K1 genes were nearly identical to Polynesian strains, although a few clustered with Australian or Taiwanese strains. These results suggest diverse origins of the Ni-Vanuatu population and raise questions about the ancient human population movements in Melanesia.

H uman herpesvirus 8 (HHV-8), or Kaposi Sarcoma Associated Herpesvirus (KSHV), is the etiologic agent of Kaposi sarcoma (KS). HHV-8 is not a widespread ubiquitous virus; its presence is mainly restricted to areas where classic or endemic KS is highly prevalent, i.e., estimates of HHV-8 seroprevalence in the general adult population range from 5% to >50% (1).
Exploiting the highly genetic variability of the HHV-8 K1 gene, molecular epidemiology led to the identifi cation of 5 major K1 subtypes (A-E), some of which appear to be strongly linked to the geographic origin of the samples. Thus, the few known subtype D strains have been reported only in inhabitants from the Western Pacifi c region (2).
For people of Oceanian ancestry (including Melanesian, Polynesian, and Micronesian), very little data are available on the clinical and molecular epidemiology of HHV-8 and its associated diseases (3)(4)(5)(6)(7)(8)(9). Thus, we studied HHV-8 in the Vanuatu, an archipelago in the Southwest Pacifi c region, formerly named New Hebrides, which contains >80 islands (6 provinces) ( Figure 1). Indigenous Melanesians, also called Ni-Vanuatu, constitute 98% of the current population of ≈210,000. A recent study suggested that HHV-8 was rare in the Ni-Vanuatu population (10).  Seropositivity was based on strict criteria, and only samples clearly reactive at a dilution >1:160 were considered HHV-8 positive. B) Age-dependent HHV-8 seroprevalence rate in 283 Ni-Vanuatu persons from 13 families originating from 4 islands (3 from Loh, 2 from Tanna, 4 from Ambae, and 4 from Esperitu Santo) of the Vanuatu archipelago. Error bars indicate 95% confi dence intervals C) Pedigrees of 2 families from Loh Island in which the presence of HHV-8 was examined in members of 3 generations. Gray circles and squares denote infected women and men, respectively. Black circles denote infected women for whom sequence of K1 gene fragment was obtained. Numbers within circles and squares indicate ages of the patients; NT, not tested.
Our goal for this cross-sectional study was to evaluate the prevalence of HHV-8 in the Vanuatu archipelago by using stringent serologic criteria and to characterize its genetic diversity.

The Study
Our work was performed on a large collection of ≈4,500 plasma and peripheral blood buffy coat (PBBC) samples from different islands of the archipelago, obtained in the framework of our previous studies on human T-cell lymphotropic virus (HTLV-1) (11,12). The fi eld survey, carried out from April 2003 through August 2005, has been extensively described (11).
A second serologic survey that used 237 plasma samples taken from 13 families with genealogic trees was performed (Figure 1, panels B and C). Among these 237 samples, 12 originated from elderly persons included in the 376 samples tested before. The HHV-8 seroprevalence was clearly age-dependent, rising from 6.9% among children 1-9 years of age to 28.2% in adults 50 years of age, followed by a new increase in persons >51 years of age (55.9%) (p<10 -4 trend χ 2 test). These results demonstrate that HHV-8 infection is endemic, widespread, and circulates in the Ni-Vanuatu population.
We then characterized these HHV-8 strains molecularly. All DNA samples (1 μg (14) and followed by a nested PCR with a second set of primers VR1S/VR2AS1 (15). All PCR products were purifi ed from gel, cloned, and sequenced. Sequences were verifi ed on both DNA strands. ORFK1 amplifi cation was obtained from 32 (21.6%) of the 148 HHV-8-seropositive samples tested but in none of the 26 HHV-8-seronegative samples. Sequences were obtained for only 30 of the 32 ORFK1-positive PCRs (Table).
Comparative sequence analysis indicates that the 30 new sequences differed from each other. Furthermore, among them, 3 groups can be clearly identifi ed. The fi rst group comprises most strains (23/30) and corresponds to the sequences found in persons from the south central islands of the archipelago (Mallicolo, Ambrym, Epi, Tongoa, Emae, and Tanna); the second group comprises 4 sequences from persons living in the northern islands of Loh (LO13, LO17, LO48) and Santo (CESW32); and the third group involves only 3 sequences (ML10, ML36, ML46) from persons living in the northern island of Motalava.
Phylogenetic analyses were performed on the 30 novel sequences obtained in this study, on all subtype D available K1 sequences, and on representatives of the different HHV-8 subtypes/subgroups, as described (4). The phylogenetic analyses were performed with all of the sequences available. These sequences include 3 strains from Japan (J24, J25, and J26), 1 from Australia (3Au1), 1 from Taiwan (TKS10), 1 from New Zealand (ZKS3), and 1 from Wallis (WalKS1) (2,4,(6)(7)(8). Our results demonstrate that the Ni-Vanuatu HHV-8 clustered in 3 different genotype D subclades, which are highly supported phylogenetically with high bootstrap values of 99% or 100% (Figure 2). The fi rst one comprising most strains corresponds to sequences closely related to each other and to the 2 Polynesian strains WalKS1 and ZKS3 (4,8). The second group (Loh/Santo), with only 4 sequences, was closely related to the Taiwanese strain TKS10 (8). The last 3 sequences from Motalava were nearly identical to the only strain from Australia (6). Furthermore, phylogenetic analysis showed a star-like tree with a long branch for the Polynesian clade, which includes most Ni-Vanuatu strains. This fi nding strongly suggests a common origin or ancestor for these strains, with a possible founder effect (Figure 2). Based on stringent serologic and molecular analyses, our study demonstrates for the fi rst time, to our knowledge, that HHV-8 infection is endemic in a Melanesian population.
Our serologic fi ndings are consistent with those found in some remote villages of Papua New Guinea (5,9). On the basis of these studies, it is tempting to suggest that intrafamilial HHV-8 transmission occurs in such populations, as previously demonstrated in highly HHV-8 endemic populations of African origin (13).
From the molecular point of view, fi nding such a high molecular diversity of HHV-8 subtype D with some Polynesian-, Taiwanese-and Australian-like strains was surprising. These heterogeneous fi ndings contrast with the more homogenous situation found for HTLV-1 genotypes in the same population (11).

Conclusions
Our molecular fi ndings suggest that HHV-8 has been introduced in the Ni-Vanuatu populations by different migrations of infected persons. This conclusion is strengthened by the clustering of the Australian-and Taiwaneselike strains in the northern islands of Loh and Motalava. A variety of scenarios have been proposed to explain the peopling of near and remote Oceania, and our data highlight the possible multiple origins of Ni-Vanuatu ancestors. Ongoing molecular studies on both viral and mitochondrial/nuclear DNA will contribute to this debate through analyses of the variations observed. Indeed, these variations are intimately linked with the dispersal of early human settlers; analyses of the genetic variability of HHV-8 can help us reconstruct the patterns of human dispersal into Oceania (11).